|
Daily Gold Bar Price Forecasting in Thailand Using Nonlinear Autoregressive Network Technique with Exogenous Factors |
|---|---|
| รหัสดีโอไอ | |
| Creator | Phimphaka Taninpong |
| Title | Daily Gold Bar Price Forecasting in Thailand Using Nonlinear Autoregressive Network Technique with Exogenous Factors |
| Contributor | Duangkamon Tupboon, Salinee Thumronglaohapun |
| Publisher | Faculty of Science, Khon Kaen University |
| Publication Year | 2569 |
| Journal Title | KKU Science Journal |
| Journal Vol. | 54 |
| Journal No. | 2 |
| Page no. | 333-343 |
| Keyword | Thailand Gold Bar Price, Gold Bar Price Forecasting, ARIMAX, NARX, Exogenous Variables |
| URL Website | https://ph01.tci-thaijo.org/index.php/KKUSciJ |
| Website title | Thai Journal Online (ThaiJO) |
| ISSN | 3027-6667 |
| Abstract | This research developed a daily gold bar price forecasting model in Thailand using the NARX model and compares it with the ARIMAX model. This study uses daily gold bar price data from 2015 to 2024, along with exogenous variables such as global gold prices, USD exchange rates, and global oil prices. The experiments were conducted by separating the dataset into three periods: 2015 - 2019, 2020 - 2024, 2015 - 2024, to evaluate forecasting performance under different levels of time-series volatility. For each period, the dataset is split in a chronological order into training dataset using 90% of the whole dataset and using the rest for testing. The results show that the NARX model outperformed the ARIMAX model in forecasting accuracy throughout all evaluation periods. This finding suggests that nonlinear models can capture complex relationships between input variables and are more suitable for forecasting gold prices in Thailand. For the gold bar price datasets spanning 2020 to 2024, which exhibited high volatility, the NARX model yielded the lower RMSE and MAPE of 486.17 and 0.93, respectively. While the forecasting of gold bar price spanning 2015 to 2019, which exhibited low volatily, the NARX model also provided the better performance with RMSE and MAPE of 144.14 and 0.36, respectively. |
